Economic growth and employment in Eastern Europe and Central Asia

Texto completo

(1)R. Soto. Growth and Employment in ECA. Economic Growth and Employment in Eastern Europe and Central Asia Raimundo Soto. Documentos de Trabajo en Estudios Asiáticos No. 11 Octubre 2014.

(2) R. Soto. Growth and Employment in ECA. Documentos de Trabajo en Estudios Asiáticos. Pontificia Universidad Católica de Chile (UC) Facultad de Historia, Geografía y Ciencia Política Programa de Estudios Asiáticos Av. Vicuña Mackenna 4860 Macul, Santiago de Chile. Contacto por correo electrónico: [email protected] (Johannes Rehner) Página web: http://www.uc.cl/icp/webcp/estudiosasiaticos/. Los resultados y opiniones presentados en el documento son de exclusiva responsabilidad de su autor y no expresan necesariamente la visión del Programa de Estudios Asiáticos de la Pontificia Universidad Católica de Chile (UC) ni de su Comité Editorial. El autor también es responsable de la originalidad del trabajo presentado. Para publicar en los Documentos de Trabajo en Estudios Asiáticos de la UC consulte la página web indicada.. Comité Editorial: Roberto Durán. Profesor del Instituto de Ciencia Política UC Marcos Jaramillo. Jefe del Programa de Estudios Asiáticos UC, profesor Facultad de Derecho UC Johannes Rehner. Profesor del Programa de Estudios Asiáticos y del Instituto de Geografía UC Raimundo Soto. Profesor del Instituto de Economía UC.

(3) R. Soto. Growth and Employment in ECA. Documentos de Trabajo en Estudios Asiáticos No. 11. Economic Growth and Employment in Eastern Europe and Central Asia 1. Raimundo Soto Instituto de Economía Centro UC de Estudios Asiáticos Pontificia Universidad Católica de Chile Octubre 2014. 1. The usual disclaimer applies Page 1.

(4) R. Soto. Growth and Employment in ECA. Index 1. Introduction ............................................................................................................................................................................................ 3 2. Comparative growth analysis ......................................................................................................................................................... 5 2.1.1. A decade lost ................................................................................................................................................................................ 8 2.1.2. A decade gained ........................................................................................................................................................................10 3. The Evolution of Foreign Trade ...................................................................................................................................................11 4. The Performance of Labor Markets ...........................................................................................................................................15 4.1. Labor supply....................................................................................................................................................................................15 4.2. Labor demand.................................................................................................................................................................................18 4.3. Labor market outcomes .............................................................................................................................................................21 4.4. Adjustment in the extensive/intensive margins .............................................................................................................25 5. Sources of Economic Growth ........................................................................................................................................................28 5.1. Methodology....................................................................................................................................................................................29 5.2. Main results .....................................................................................................................................................................................30 6. The Determinants of Employment .............................................................................................................................................32 6.1. A simple model of the labor market .....................................................................................................................................32 6.2. Decomposing employment growth .......................................................................................................................................36 7. Conclusions and Stylized facts ......................................................................................................................................................39 References ........................................................................................................................................................................................................43 Appendix 1: List of countries used in International Comparisons. ...................................................................................44 Appendix 2: Computing Sources of Growth .................................................................................................................................45 Appendix 3: A model of the labor market .....................................................................................................................................46 Appendix Table 1: Data Sources and Coverage ..........................................................................................................................47 Table 1 Table 2 Table 3 Table 4 Table 5 Table 6 Table 7 Table 8 Table 9 Table 10 Table 11 Table 12. GDP per capita in ECA and other regions of the world in 2010 Main Performance Indicators in ECA and other regions of the world Foreign Trade: Export and Import Values Foreign Trade: Trade Structure, 2010 Main Demographic Indicators Total Employment, Public sector Employment, and Sectorial Employment Average Labor Productivity, Real Wages, and Hours Worked Unemployment Rates, selected years (%) Adjustments in the Extensive and Intensive Margins of the Labor market Sources of economic growth, 2000-2011 Econometric estimation of labor demand models. Contribution of fundamentals to the cumulative change in employment. Figure 1 Figure 2 Figure 3. Crisis and recovery in ECA Economic Prosperity in Estern Europe and Central Asia Average labor productivity and real wages. Page 2.

(5) R. Soto. Growth and Employment in ECA. 1. Introduction The economic development of Eastern Europa & Central Asia (ECA) countries in the last two decades is marked by the transition from a planned economy to a market economy and their progressive re-insertion in the world economic community. The path has not been easy for most countries as the physical and social cost of the massive restructuring of their economies have been dire. In the early 1990s all countries suffered from collapsing economic activity, industrial and financial disarray, and soaring unemployment. Some economies suffered also from high inflation. The 2000s, on the contrary, have been years of recovery in all economies and outright bonanza in a few of them. Countries benefited from the discipline brought upon firms by increasing market competition as a result of foreign trade opening, market de-regulation, and privatization while the retrenching of the government from productive activities gave way to individual initiative and creativity. Some countries also benefited from ample external funds in the form of foreign direct investment. In this generally positive outlook ECA economies have transformed and benefited from rejuvenation in varying degrees depending on their initial backwardness, the wisdom of the policies implemented and, to a lesser degree, their good fortune in terms of natural resources, location, institutional fabric and other determinants of their insertion in global markets. Massive economic restructuring is a costly, lengthy and risky process: resources have to be relocated, non-profitable businesses have to be terminated, and emerging opportunities seized. Institutions have to be adjusted or created altogether to meet new, more challenging demands in efficient, non-corrupt manners. Political capacity is also needed to maneuver promptly and swiftly to ameliorate the social costs of restructuring for some groups in society and secure the benefits for the majority of the population. The policy responses to the challenges of modernization in ECA has been mixed and their fate also heterogeneous. Some economies embraced economy wide, far reaching reforms and transformations, while others advanced rather timidly. Some ECA economies have been quite successful in this process while others have struggled for a long time to return to shape. Shocks and transformations elsewhere in the world have also affected this process. High commodity prices have played a significant role in fostering growth in the last decade, in particular the windfall received by resource-rich ECA countries exporting oil, gold or diamonds. The emergence of East Asia and China as leading trade partners of the developed economies increased external competition for ECA countries in semi-manufactured consumer and investment goods, although they also opened the door to substantially large markets. Finally, a buoyant international financial market provided ample access to resources to ECA for both investment and market development. The recent global crisis put a transitory stop to the golden decade of the 2000s. Financial flows dried up, commodity prices collapsed and external demand plummeted. While globally the gross domestic product contraction was about two percent between 2008 and 2009, in ECA it was. Page 3.

(6) R. Soto. Growth and Employment in ECA. more than four percent. In 2009, output in 20 of 29 Eastern European and Central Asian economies2 contracted and only Central Asian countries appeared to have been spared from the downturn. However, such contraction in activity proved to be short-lived: by 2011 GDP had regained positive growth in all but one country (Slovenia) and (unweight) regional output growth had reached 4.5% year on year basis. Although the setback in activity appears to be a transitory phenomenon, employment was disproportionally affected by the crisis. Between 2008 and 2009 employment declined in 18 of the 29 economies. Inasmuch as history repeats itself, the response of ECA economies to previous suggests that the excess sensitivity to economic downturns is a defining trait of ECA’s labor markets and that recovery in employment take long time to materialize. By 2010, employment was around 5% below its pre-crisis level in 9 economies and unemployment rates were on the rise. The recent global downturn proved the vulnerability of small open economies in Eastern Europe and Central Asia to foreign shocks. This, of course, is a well-learned lesson from previous crisis in other regions of the world. While the upswing of open economies is sweet, the downswing tends to quite painful. Much of the harshness of the downswings depends on the ability of countries to avoid the build-up of imbalances prior to the crisis and to adjust quickly to changing environments during the downturn. A poorly operating labor market can play a crucial role in deepening the crisis, while a well-oiled one can soften the costs by allowing changes in the intensive margin (intensity of use of workers) to cushion against the otherwise inevitable adjustment in the extensive margin (layoffs). In this regard, ECA economies portray precisely this tension: a number of countries entered the crisis with high and possibly chronic unemployment levels (e.g., the Balkans and the smaller CIS3 economies) while others had already achieved single-digit unemployment rates by the mid-2000s. In almost all economies unemployment increased as a result of the crisis but in those where adjustments in the intensive margin were significant, unemployment rates are returning to pre-crisis levels faster. This paper reviews the long-run growth and employment experience of ECA economies in the last two decades. The task is ambitious as there is a remarkable heterogeneity among these countries in initial conditions, the quality of the implemented policies and reforms, and their economic outcomes. Therefore the goals are three. First, to identify the stylized facts characterizing the evolution of the different economies from a macroeconomic viewpoint and in terms of comparative long-run growth (in section 2), the evolution of foreign trade (section 3), the performance of labor markets (section 4) and the sources of economic growth (section 5). Second, to study the determinants of employment in the period 2000-2011 and evaluate the role in creating employment of sustained economic growth, changes in real wages, the retrenchment in public employment and the concomitant decline in public employment, and the expansion in foreign trade. Econometric evidence on this is presented in section 6. Section 7 provides the conclusions and main stylized facts, while the most technical aspects are relegated to a set of appendices.. 2 3. Kosovo is excluded from the analysis due to lack of data. Commonwealth of Independent States Page 4.

(7) R. Soto. Growth and Employment in ECA. 2. Comparative growth analysis When analyzing the economic performance of ECA countries it is important to benchmark them against the rest of the economies in the world. This allows controlling for cross-country international phenomena and transient shocks. I consider first real GDP per capita at PPP prices as an encompassing indicator of economic development as well as a proxy of the welfare levels of the population.4 Naturally, equalizing welfare with GDP per capita ignores issues derived from the very diverse income distributions that characterize countries in different regions and suffer from the limitations of national accounts, but it allows for easier comparison across countries and in time. In Table 1 I present data for around 187 economies of the world, grouped by geographic area. I use simple averages (non-weighted) for regional comparisons because in ECA, as well as other regions, large economies tend to dominate regional averages and trends. In ECA, Russia and Turkey would dominate weighted averages while in LAC it would be Brazil, in South Asia it would be India, and in East Asia, China. I acknowledge that there is a significant heterogeneity among ECA countries and, consequently, undertake a more detailed country-by-country analysis below. It can be seen that as of 2010, ECA countries had on average higher per-capita income than any other region in the world, excepting of course the developed economies. Average income levels, however, were still below OECD and EU standards by a significant margin. In fact, the highest income economy in ECA –the Czech Republic—would barely make it to the lowest income segment of the developed economies group, while the poorest country –Tajikistan—has income levels comparable to low-income countries in Middle East & North Africa (MENA) or South Asia. The evidence indicates furthermore that the heterogeneity (variance) in terms of income, and thereby development levels, in ECA is not significantly different than that in MENA, LAC, or East Asia & Pacific (EAP) –when excluding Hong Kong and Singapore from the latter.. Table 1 GDP per capita in ECA and other regions of the world in 2010 Constant US$ of 2005 Regions5 Average Lowest Highest Europe & Central Asia 10,976 1,886 24,833 Middle East & North Africa 7,836 2,023 26,183 Latin America & Caribbean 9,227 999 23,041 South Asia 3,256 1,079 5,168 East Asia & Pacific 9,354 832 51,326 Sub-Saharan Africa 3,637 303 31,471 Developed economies 34,385 21,660 71,048 World 11,360 303 71,048 Source: Own elaboration based on World Development Indicators 2012, World Bank.. 4. GDP per capita based on purchasing power parity (PPP) is GDP converted to international dollars using purchasing power parity rates. An international dollar has the same purchasing power over GDP as the U.S. dollar has in the United States. Data are in constant 2005 international dollars. 5 The list of countries is in the appendix. Page 5.

(8) R. Soto. Growth and Employment in ECA. While the development levels of ECA economies may be similar to those in other emerging regions of the world, their evolution in the last two decades is markedly different. As shown in Table 2, economic growth has gone through two very dissimilar phases. First, a significant contraction in economic activity in the early 1990s as a result of the collapse of the Soviet Union, political unrest and civil wars, and the replacement of the socialist command economy –with varying degrees of intensity and success- by a private sector, market-driven economic system. For most economies in the region the 1990s amounted to a decade lost in terms of economic growth. Between 1990 and 1994 GDP in ECA declined on average by around 35%. The second half of the decade saw a few ECA countries recover swiftly while most other struggle to improve. On average economic growth was a modest 2.9% per year, allowing for a mild recovery of ECA’s average GDP to around 80% of its level of 1990. The poor performance of ECA countries in the 1990s contrasts with the buoyancy of most other regions and the world economy, particularly in the second half of the 1990s. Even lagging regions –such as Sub-Sahara Africa—display a better performance. The second phase is one of sustained economic growth for most ECA economies, although the initial rapid growth of the early 2000s seems to be yielding in recent years to more moderate expansions in economic activity. The majority of ECA countries have already recovered the precrisis income levels indicating that the first stage of the transition towards modern market economies has been completed. In fact, during the early 2000s ECA was the fastest growing region in the world and there was a significant catch-up with the mature economies of the European Union. In general terms the 2000s can be characterized as a decade gained. Catching up in ECA was further fueled by differences in population growth: while in the developed economies population grows slowly but systematically, in ECA average growth rates are close to zero or even slightly negative. This is a phenomenon unique to ECA as in other regions percapita GDP growth is significantly lower than GDP growth (by around one and a half percentage point at the world level, as shown in Table 2). Of course, this raises issues of manpower availability, on one hand, and population ageing and obsolescence, on the other, which I discuss below. As explored below, ECA economies are notoriously heterogeneous in terms of population and economy size, as well as their endowment of natural resources. A few countries (including Azerbaijan, Kazakhstan, Turkmenistan and Russia) enjoy significant deposits of hydrocarbons and valuable minerals (gold) and, therefore, benefited from the commodity price boom prior to the recent global crisis. As shown in Table 2, this windfall is sizable as, on average, natural resource rents6 in ECA are quite high for world standards, second only to oil-rich Arab nations in MENA. Furthermore, it can also be noted that between the 1990s and 2000s resource rents in ECA as share of GDP expanded by around four percentage points, indicating the ample magnitude of the windfall. In the last two decades ECA countries underwent a massive restructuring of their economies from state-driven socialism to private sector-led market economies. Naturally, countries have engaged in reforms with varying levels of enthusiasm and wisdom, and while some economies have enjoyed substantial success others continue to struggle. One key indicator of the degree of 6. Rents are defined as the volume of production times the difference between the international price and the cost of production. Page 6.

(9) R. Soto. Growth and Employment in ECA. success of reforms is labor productivity: if reforms are successful, labor productivity should increase on a sustainable basis and eventually converge to the standards of highly developed economies. I use real GDP per worker as a crude measure of labor productivity. It can be seen that as of the early 1990s labor productivity in ECA was well below international standards (one third of the OECD level and 15% below LAC averages), indicating a significant backwardness in the region even if we take into account that these were years of political and economic turmoil. In the subsequent recovery years, labor productivity has grown systematically –particularly in the 2000s—to achieve around 70% of the EU and leveling off with LAC and EAP.. Table 2 Main Performance Indicators in ECA and other regions of the world (simple averages) Europe & Central Asia. Middle East & North Latin America Africa & Caribbean South Asia. East Asia & Pacific. SubSaharan Africa. Developed Countries. World. 0.1 4.8 4.2 4.7. 2.1 3.5 2.6 1.4. 1.0 3.9 4.0 4.1. -2.3 2.1 1.7 2.4. 1.3 2.9 1.9 0.6. -0.7 2.4 2.5 2.7. 9.3 9.3 10.2 12.6. 1.7 1.3 1.3 1.6. 7.7 6.8 9.0 11.2. 37.1 40.6 43.8 46.3 45. 16.2 17.3 18.9 21.3 187. Real GDP growth (annual average, percent) 1990-1994 1995-1999 2000-2004 2005-2011. -9.6 2.9 6.1 4.4. 5.6 4.6 3.6 5.5. 3.2 3.4 2.6 3.3. 4.8 5.3 5.6 6.3. 4.9 4.1 4.0 4.5. Growth in real GDP per capita (annual average, percent) 1990-1994 1995-1999 2000-2004 2005-2011. -9.9 3.0 6.1 4.1. 2.2 2.0 1.2 3.0. 1.7 2.0 1.3 2.1. 3.3 3.6 3.7 4.6. 3.0 2.4 2.6 3.1. Total natural resources rents (percent of GDP) 1990-1994 1995-1999 2000-2004 2005-2011. 8.7 7.1 12.3 12.6. 18.7 18.2 25.7 32.5. 5.2 3.5 4.7 7.5. 5.9 4.7 3.9 4.6. 6.9 5.5 6.5 6.8. GDP per person employed (thousands of US$ at PPP of 1990) 1990-1994 1995-1999 2000-2004 2005-2011 Countries. 12.0 12,0 15.4 20.0 29. 15.2 14.7 15.3 16.7 10. 15.3 16.6 171 19.2 30. 5.4 6.2 6.8 8.1 7. 12.9 15.3 17.2 20.3 22. 2.4 2.4 2.5 2.9 44. Source: Own elaboration based on World Development Indicators 2012, World Bank.. This portrait of ECA economies vis-à-vis other regions of the world is informative of general trends but hides the very dissimilar nature of ECA countries as well as their complex interactions. The following sub-sections provide an in-depth analysis of the experiences of countries belonging to the CIS, the Balkan region, and those in the European Union to which I add Croatia and Turkey (EU+). Arguably, the latter constitutes a very different case from the rest of ECA but serves as benchmark for contrasting their experiences.. Page 7.

(10) R. Soto. Growth and Employment in ECA. 2.1.1. A decade lost The 1990s constitute a decade lost for economic growth in ECA.7 All economies, with the only exception of Poland and Slovenia, either contracted or stagnated in real terms. While growth in Poland was above 3% per year and helped reduce the income gap vis-à-vis developed economies, in Slovenia it was only a modest 1.5% per annum, insufficient for any catch up with European countries. Bosnia and Herzegovina grew very fast after the end of the 1992-1995 war, but it is estimated that by 1999, income was still well below the 1990 levels, thereby also falling in the lost-decade group.8 In this generalized negative scenario there are remarkable differences in economic performance among ECA economies in the 1990s. While for most countries the decade initiated with massive recessions, the depth and length of such downturns and the strength of their recoveries are remarkable different, as shown in Figure 1. The economies in EU+ group and the Balkans experienced deep but short-lived recessions with GDP contracting typically by around 20% to 30% in a period of three to four years, after which economic growth resumed. On the contrary, the downturns in the CIS economies lasted much longer (from five to seven years), were more profound (with cumulated declines above 50%) and were not necessarily followed by vigorous growth. The cumulative decline in GDP in the EU+ countries was moderate when compared to the CIS economies (although quite large for the standards of international business cycles9). On average it was around 25% and some economies it did not reach 15%, as in Estonia, Hungary, Poland, and Slovenia. Only in Latvia and Lithuania, GDP contractions of 45% approached the levels of the CIS economies. Despite the mildness of the negative shock, recovery patterns have been quite heterogeneous. Some countries recovered quite fast –in less than five years—even if they had suffered massive contractions (e.g., the Czech Republic with an initial drop of 35% of GDP). Others endured slow recoveries for around a decade before reaching the pre-crisis GDP levels (e.g., Bulgaria, Romania) even if the initial negative shock was not significantly high (e.g., Estonia). As discussed below, the differences in recovery patterns are linked to significant differences in total factor productivity gains and the working of the labor market.. 7 Turkey is an exception to ECA’s performance in that in 1990 it was already a relatively developed market economy, it did not undertake massive restructuring, and did not experience a severe depression. I therefore use it mainly as a benchmark. 8 IMF (1998) also casts doubts on the quality of the national accounts of Bosnia and Herzegovina in the 1990s. 9 These episodes qualify as “great depressions” according to Kehoe and Prescott’s (2007) definition: “To be a great depression, a negative deviation from trend must satisfy two conditions. First, it must be a sufficiently large deviation. Our working definition is that a great depression is a deviation of at least 20 percent below trend. Second, the deviation must occur rapidly. Our working definition is that de-trended output per working-age person must fall at least 15 percent within the first decade of the depression.”. Page 8.

(11) R. Soto. Growth and Employment in ECA. Figure 1 Crisis and recovery in Eastern Europe and Central Asia. Note: The “cumulated decline in GDP” measures the percent drop in GDP between 1990 and its lowest level achieved afterwards. The “years to recover” measures the number of years that passed between the lowest GDP level and its return to the 1990 level. Countries in red have not yet regained the 1990 GDP level: plotted values are the expected date of recovery based on current growth levels.. Source: Own elaboration based on World Development Indicators 2012, World Bank.. The CIS economies, on the other hand, suffered very deep, long lasting downturns. In all economies –except Uzbekistan—GDP contracted by more than 35%. In several countries economic growth after the crisis was not vigorous and, given the magnitude of the initial downturn, recoveries took much longer than in the EU+ group (around 11 years on average). Even as of today in some economies GDP has not yet regained pre-crisis levels (e.g., Ukraine, Moldova, and Georgia). Again, the hesitant recovery of the 1990s in the CIS economies is linked to productivity issues as discussed below.. Page 9.

(12) R. Soto. Growth and Employment in ECA. 2.1.2. A decade gained While the 1990s were the years of collapse, the 2000s have been years of bonanza for most countries in ECA. As shown in Figure 2, in general lines economic growth has followed the classical convergence pattern whereby countries with initial lower levels of income per-capita tended to grow faster and to some extent catch-up with more advanced economies. As evidenced by the location of countries along the negatively-sloped regression line, the CIS economies were on average far behind the EU+ group in terms of income level by the end of the 1990s and have subsequently grown much faster. Other reasons, such as natural resource rents, have also been instrumental as discussed below. The EU+ economies have grown in the 2000s at an average annual rate of 3.3% and in a very homogenous fashion: the fastest growth economy being Lithuania (at 4%) and the slowest being Hungary (at 1.8%). In the CIS and the Balkans, on the contrary, heterogeneity is the norm. The fastest economic growth is observed in Azerbaijan where gas-exports have fueled double-digit growth rates (at an annual rate of 13.5% on average). Lower but still quite sizable growth rates have been achieved by Armenia, Belarus and Kazakhstan. Underperformers in this group are clearly Macedonia and Kyrgyzstan. Albania, Montenegro and Serbia locate midway between CIS and EU+ economies.. Figure 2 Economic Prosperity in Eastern Europe and Central Asia. Source: Own elaboration based on World Development Indicators 2012, World Bank.. Page 10.

(13) R. Soto. Growth and Employment in ECA. 3. The Evolution of Foreign Trade One of the most important aspects of the transition from socialist rule to market economy is the opening to foreign trade and the intensification of market competition. Several countries literally abandoned trade autarky whilst others that had previously traded with the global economy saw themselves competing without the support of the government in terms of subsidized exchange rates, forced trading with members of the political coalition, and soft-budget constraints that would allow surviving despite systematic losses.10 Among other things, foreign competition has increasingly brought discipline to ECA economies, it has demanded improvements in the quality of the goods sold and in the delivery conditions, it forced firms to become more efficient and required exporters to meet environmental standards. Some countries have been more successful than others and some industries within countries have been less apt to adapt than others. Foreign trade has undoubtedly become a key aspect of economic development for most ECA economies. As shown in Table 3, foreign trade has increased markedly during the past two decades in most economies. First, note that on average for all countries, imports are significantly larger than exports, i.e., ECA has a systematic trade deficit. Such deficits are financed in the long run by surpluses in the services accounts; in the short-run capital flows can be used to finance but considering the size of such deficits the latter option would amount to accumulating unsustainable debt levels. Historically such trade deficits would have been a source of concern but nowadays trading in services has become a major industry: for example, several Eastern European economies have well developed tourism facilities that provide significant returns to the economy. Second, note that larger economies tend to trade less than smaller economies. Trade in Russia, Turkey and Poland is significantly less than in smaller economies with similar levels of development such as Slovenia or the Czech Republic. This is a worldwide empirical regularity as documented in the literature on “gravity models of trade” which I discuss below: large-size economies can rely on domestic markets for a stable demand on which to base the development of industries and, more often than not, they count on the natural resources contained in extensive geographical areas. Economies, particularly in the western regions of ECA, are small in both size and population. Third, the countries in the EU+ group tend to trade significantly more than those in the CIS or the Balkans. Exports by EU+ economies in the period 2000-2011 are ten percentage points of GDP higher than those of the CIS (which include energy exporters) and twenty percentage points higher than the Balkan countries. Naturally, accessing the EU provides ample opportunities to sell products in a very large and rich trade area and can explain the leadership of EU+ economies in trade. However, in order to use such opportunities producers ought to be as efficient as the “domestic” producers; therefore, the higher share of exports in EU+ countries also indicates the higher competitiveness levels. 10. For a detailed analysis see the companion paper in Soto (2014). Page 11.

(14) R. Soto. Growth and Employment in ECA. Fourth and related to the previous issue, it can be seen that exports as share of GDP have increased quite significantly in the EU+ group and the Balkans but have stagnated in the CIS. The same is observed in terms of imports. As discussed below, the very different structures in terms of trade partners in ECA could provide an explanation for this regularity. Fifth, as is the case with most markets in ECA, there is substantial heterogeneity among countries vis-à-vis trade volumes. In the last decade, a large number of economies exported around 30% to 40% of GDP, whilst very few surpassed the 50% mark typically EU+ countries.. Table 3 Foreign Trade: Export and Import Values Exports (as % of GDP) 1990-2000 2001-2011 Balkans Albania 13.0 25.2 Bosnia 24.4 34.4 Macedonia 38.0 44.4 Montenegro 36.8 38.2 Serbia 20.4 27.8 Average 26.5 34.0 CIS Armenia 30.2 23.9 Azerbaijan 39.5 55.1 Belarus 56.8 61.6 Georgia 29.2 31.3 Kazakhstan 43.1 48.5 Kyrgyzstan 35.1 46.2 Moldova 44.4 46.6 Russia 32.6 32.5 Tajikistan 53.2 35.1 Turkmenistan 66.1 65.1 Ukraine 39.1 51.7 Uzbekistan 26.3 35.8 Average 41.3 44.4 EU members, Croatia and Turkey Bulgaria 49.3 53.2 Croatia 45.4 41.0 Czech Republic 49.3 64.4 Estonia 71.9 74.4 Hungary 45.7 74.8 Latvia 49.8 45.8 Lithuania 47.3 57.9 Poland 23.5 36.6 Romania 25.3 31.2 Slovak Republic 56.1 78.8 Slovenia 59.8 62.1 Turkey 18.5 23.2 Average 45.2 53.6 Source: own elaboration based on World Bank database.. Imports (as % of GDP) 1990-2000. 2001-2011. 41.0 83.2 47.5 51.1 30.9 50.7. 48.9 70.8 63.7 67.4 48.4 59.8. 57.4 47.9 60.9 50.5 45.6 47.5 56.6 25.3 59.0 67.5 39.2 30.6 49.0. 44.2 39.6 67.7 50.4 39.5 67.3 84.8 22.2 65.8 52.7 52.5 32.0 51.6. 50.5 50.3 49.8 79.9 47.3 51.0 53.5 24.5 31.5 60.7 58.1 21.5 48.2. 64.5 46.5 62.7 77.8 74.2 57.0 64.3 38.9 39.7 81.7 62.5 26.0 58.0. Page 12.

(15) R. Soto. Growth and Employment in ECA. These stylized facts ought to be linked with the trade structure of the countries in terms of goods exported and imported as well as the trade partners of each country for several reasons. On one hand, elements such as geographical proximity, historical and cultural ties, and trade agreements usually have influence on the type of goods traded and the origin and destination of such trade. On the other hand, the endowment of natural resources and the relative availability of production factors (manpower, land and capital) also shape the nature of trade among countries. In the case of ECA the potentially overwhelming influence of Russia and the EU can also affect trade patterns as smaller economies tend to gravitate around large-size countries. In Table 4 I provide a summary of trade structures –in partners and goods—to identify stylized facts. The results are quite clear. It can be seen that the EU is the major trading partner of all economies except a few countries in the CIS group for which Russia is the main source of imports and an important destination for exports. The importance of EU is, not surprisingly, the largest in those countries that have joined the trade union: around 60% of exports are directed to the EU from where it comes 55% of imports. However, note that Balkan countries, which are not part of the EU, trade only slightly less with around 50% of total export and imports linked to the EU. Indeed, and more strikingly, while in the CIS trade with the EU is not as important as elsewhere, the main trade partner of Russia is the EU: around 50% of all Russian imports come from the EU and, even when excluding energy, almost 40% of Russian exports go to the EU. Russia, on the other hand, is not a major trade partner of the Balkans or the EU+ countries both in terms of a destination for exports and as a source of imports. On average less than 5% of export volumes and 10% of import volumes correspond to trading with Russia. Some EU+ economies are more dependent on Russia’s exports: the Baltic countries import up to 20% of total imports from their giant neighbor. Russia, nevertheless, is the major trade partner for the smaller economies of the CIS with only a few significant exceptions. As a destination for exports, some economies are quite integrated with Russia: Belarus, Kyrgyzstan, Moldova, Ukraine and Uzbekistan sell over 20% of total exports to the Russian markets, a figure that increases significantly if energy transactions are excluded. These economies, particularly Belarus, also import a sizable proportion of their foreign purchases from Russia. For other economies such as Armenia, Azerbaijan, Kazakhstan, Moldova and Tajikistan the Russian market is important but not dominant. Finally, for Georgia and Turkmenistan Russia represents a small share of their trade volumes. Beyond identifying stylized facts, the importance of scrutinizing the structure of trade in terms of partners and goods lies in determining the diversification of ECA economies: trade diversification is an important mechanism to reduce the risk inherent to participating in international markets. The risk can be split in two components. First, the lack of diversification in terms of exported goods, whereby countries that concentrate on a few goods (in the limit, monoexporters) tend to pass-on to the economy the wide fluctuations that characterize commodity prices. Second, the lack of diversification in terms of partners, whereby economies tend to become dependent on the business cycles of another country (e.g., Russia or the EU).. Page 13.

(16) R. Soto. Growth and Employment in ECA. Table 4 Foreign Trade: Goods and Partner Structures, 2010 Exports to (as % of total exports). Non-Energy Exports to (as % of total non-energy exports) European Russia Union. Russia. European Union. 56.9 43.9 58.5 45.1 51.1. 3.4 1.9 7.2 9.3 5.5. 62.0 36.4 43.4 43.8 46.4. 14.3 30.2 54.4 7.1 24.6 34.3 26.4 17.6 8.9 24.3 30.0 24.7. 51.8 12.3 10.4 26.4 26.9 6.9 27.1 37.0 40.8 12.3 19.4 15.6 23.9. 18.7 18.2 62.7 10.7 34.7 20.0 12.0 25.3 17.4 26.5 24.9 24.6. 33.3 34.9 18.1 29.1 28.4 8.3 27.6 52.1 6.8 27.4 30.8 16.7 21.1. EU members, Croatia and Turkey Bulgaria 2.4 55.8 2.7 Croatia 1.5 52.1 1.6 Czech Rep. 2.3 70.3 2.3 Estonia 8.8 61.8 9.8 Hungary 3.2 67.0 3.2 Lithuania 10.3 49.8 13.3 Latvia 7.2 52.1 8.4 Poland 4.4 69.8 4.4 Romania 1.7 62.8 1.8 Slovak Rep. 3.1 62.8 3.2 Slovenia 3.6 61.7 3.6 Turkey 4.1 58.5 4.2 Average 4.4 60.4 4.9 Source: own elaboration based on UN TRADECOM database.. 60.3 55.7 70.5 60.5 67.3 49.2 49.9 69.8 66.6 63.7 61.9 59.6 61.3. 9.9 7.2 4.3 12.0 6.2 22.4 15.1 8.6 5.3 8.5 1.1 11.9 9.4. 50.0 56.2 65.8 55.4 63.0 44.5 41.2 65.8 58.1 47.6 72.5 50.5 55.9. Balkans Albania Bosnia Macedonia Serbia Average. Russia. European Union. 0.5 0.3 1.3 5.4 1.9. 58.1 43.8 57.1 44.4 50.9. 0.6 0.3 1.3 5.4 1.9. 14.3 2.7 37.2 5.5 10.3 33.6 26.3 17.6 1.4 23.7 23.6 17.8. 52.1 67.4 27.7 32.5 53.5 7.1 27.0 49.4 40.7 10.8 20.4 12.7 33.4. Imports from (as % of total imports). CIS Armenia Azerbaijan Belarus Georgia Kazakhstan Kyrgyzstan Moldova Russia Tajikistan Turkmenistan Ukraine Uzbekistan Average. Page 14.

(17) R. Soto. Growth and Employment in ECA. 4. The Performance of Labor Markets While vigorous economic growth and recovery seems to be a general characteristic of the region in the 2000s, the performance of the labor markets is distinct and very heterogeneous. Heterogeneity comes from both sides of the labor market. On one hand, there are important differences in the behavior of the labor supply in terms of population growth and participation rates. I review these issues first. On the other hand, there are also noticeable differences in labor demand in terms of the sectorial distribution of workers and the role of the public employment. I review these issues in the second part of this subsection. Naturally, this heterogeneity produces quite different outcomes in the labor markets of the economies of ECA. Unfortunately the data on labor market outcomes are quite restricted and frequently of weak quality. In the third part I therefore focus on three market outcomes for which there is some systematic information: average labor productivity, real wages, hours worked, and unemployment. Information for other market outcomes–such as underemployment—is largely missing (particularly for CIS countries). Data are occasionally distorted for political reasons (e.g., official unemployment rates systematically below 1% in some Central Asia countries).. 4.1.. Labor supply. The supply side of the labor market depends largely on the size and growth of the population and the participation rate of the economically active population. The latter is customarily defined as the ratio of the population that participates in the labor force and the overall size of their cohort (national population of the same age range), usually those between 15 and 65 years of age. These variables do not respond to short-term fluctuations in the economy, moving slowly in time and largely in response to structural changes in society. I therefore treat them as exogenous forces in the analysis of the labor market performance in ECA; considering our interest in analyzing the last decade this does not seem to be very restrictive. Population in many ECA countries is stationary; in the period 1990-2011 total population has remained virtually unchanged in Albania, the Czech Republic, Hungary, Kazakhstan, Moldova, Montenegro, Poland, the Russian Federation, Serbia, the Slovak Republic, and Slovenia. On the other hand, it has declined by more than five percent in another eleven countries (Armenia, Bulgaria, Bosnia and Herzegovina, Belarus, Croatia, Estonia, Georgia, Latvia, Lithuania, Romania, and Ukraine). In only seven countries population increased more than five percent between 1990 and 2010: Azerbaijan, Kyrgyzstan, Macedonia, Tajikistan, Turkey, Turkmenistan and Uzbekistan. When compared to other regions in the world, it can be seen that population growth in ECA economies is extremely slow for any standard. In fact, it is even slower than in the demographically mature economies of the Eurozone. Only in Turkish countries population grows at the world level. This phenomenon determines, to some extent, the development path of ECA countries as they cannot count on an expanding population to provide additional manpower in the labor market. Page 15.

(18) R. Soto. Growth and Employment in ECA. A stationary or declining population indicates that countries have achieved some demographic maturity and suggests also that adjustment in the labor market would most likely take place in the intensive margin (higher use of existing resources) and not in the extensive margin (increase in resources). I return to this issue below. In addition to very stationary population, ECA countries are characterized by the evolution of participation rates that do not contribute to the extensive margin. The data in Table 5 indicate that in ten countries participation rates have either remained practically stagnant or declined slightly (Azerbaijan, Bosnia and Herzegovina, Hungary, Kyrgyzstan, Kazakhstan, Latvia, Macedonia, Russia, Turkmenistan, and Uzbekistan). Furthermore, participation rates declined in 16 economies, ranging from Tajikistan (where the decline has been around one percentage point) and Moldova (with a drop of 14 percentage points). In only in Slovenia participation rates have expanded in the last 20 years by around two percentage points. The general trend towards lower participation rates may be the result of several forces at work usually dubbed “push” and “pull” factors (see Becker, 1965). In most market economies, evidence suggests that participation changes with phenomena such as the entry/exit of women from the labor force, changes in retirement periods and pension levels, expansions and contractions in the informal sector, and the abolishment of discriminatory practices against women. It should be noted, however, that a large number of countries in ECA have higher participation rates than the European standard and, therefore, there may be a tendency in these countries to converge the Eurozone as a result of the homogenization in working conditions, particularly in those countries that are already member of the EU (e.g., Czech Republic, Estonia, Lithuania, Latvia, Romania, Slovak Republic). The combination of stationary population and stable or declining participation rates indicate that the supply of national workers tends to be quite stable and that labor-market conditions may be primarily determined by labor demand. In fact, the labor force has remained stagnant or declined in the last 20 years in 17 ECA countries. Considering the size of the downturns in the early 1990s and the reputedly high education levels of the labor force, it comes as no surprise that ECA countries tend to show very significant levels of out-migration. The lowering of migratory barriers by Western European economies in the 1990s facilitated migration. In the last two columns of Table 5 I collected data on the stock of nationals living abroad and its share in national population. It can be seen that in only three economies (Czech Republic, Poland and Russia) there is net immigration (i.e., the number of nationals living abroad is smaller than the number of foreigners living in the country) while in another two economies migration is not an issue (Estonia and Slovenia). In all other ECA countries outmigration is significant, being extremely high in Balkan countries (Albania, Bosnia and Herzegovina, and Macedonia), quite high in some CIS countries (Armenia, Azerbaijan, Georgia and Moldova) and comparatively moderate in the economies of EU+ group. The gross out-migration in some countries is compensated by a significant inflow of migrants –e.g., Belarus, Russia—indicating an important labor mismatch.. Page 16.

(19) R. Soto. Growth and Employment in ECA. Table 5 Main Demographic Indicators Population Growth Rate (annual average, %). Participation Rates (average, %). Population outside country in 2010 Share of Millions Population (%). 1990-2000. 2001-2011. 1990-2000. 2001-2011. -0.7 -1.5 0.5 -0.1 -0.5. 0.3 0.2 0.2 -0.3 0.1. 63.4 45.1 54.3 54.3. 60.7 44.8 53.9 62.4 55.5. 1.35 1.45 0.33 0.20 0.83. 42.1 38.6 16.1 2.7 24.9. -1.4 1.2 -0.2 -0.8 -0.9 1.1 -0.2 -0.1 1.5 2.1 -0.5 1.9 Average 0.3 EU members, Croatia and Turkey Bulgaria -0.8 Croatia -0.8 Czech Rep. -0.1 Estonia -1.4 Hungary -0.2 Lithuania -0.6 Latvia -1.2 Poland 0.1 Romania -0.3 Slovak Rep. 0.2 Slovenia 0.0 Turkey 1.7 Average -0.3. 0.0 0.9 -0.3 0.1 0.8 0.8 -0.2 -0.3 1.2 1.3 -0.7 1.2 0.4. 66.2 63.3 62.5 65.7 69.7 65.6 63.7 62.1 66.9 60.2 60.2 59.3 63.8. 60.1 63.8 56.4 64.1 70.1 65.1 49.0 61.2 65.6 60.6 57.8 60.0 61.2. 0.55 1.33 0.68 0.90 0.65 0.41 0.39 -0.76 0.58 0.07 1.59 1.12 0.63. 17.9 14.6 7.2 20.1 4.0 7.5 10.9 -0.5 8.4 1.4 3.5 4.0 8.3. -0.6 0.0 0.2 -0.2 -0.2 -0.5 -0.5 -0.1 -0.4 0.1 0.3 1.2 -0.1. 55.5 55.5 61.1 63.0 50.8 63.1 61.9 59.1 62.7 61.7 56.4 54.2 58.8. 52.6 53.0 59.2 59.8 49.9 57.7 59.1 54.9 56.9 59.7 58.7 48.1 55.8. 1.10 0.17 -0.06 0.00 0.18 0.30 -0.06 2.35 2.66 0.39 0.00 2.99 0.84. 14.6 3.7 -0.6 0.0 1.8 9.1 -2.5 6.2 12.4 7.2 0.0 4.1 4.7. Euro Area 0.3 0.5 54.5 56.2 22.2 Latin America 1.6 1.2 63.0 65.3 30.5 World 1.5 1.2 65.9 64.9 193.6 Source: Own elaboration based on World Bank (2011) and Migration and Remittances Factbook 2011.. 6.7 5.1 2.8. Balkans Albania Bosnia Macedonia Serbia Average CIS Armenia Azerbaijan Belarus Georgia Kazakhstan Kyrgyzstan Moldova Russia Tajikistan Turkmenistan Ukraine Uzbekistan. There are three main explanations for these strong migration currents. First, prior to 1990, migration was severely limited by national authorities. As shown in Table 5, migration is a significant in the world and in the Eurozone or other emerging economies. Therefore, part of the increase in the stock of national population living abroad may reflect a simple convergence to world standards. In addition, some migratory flows observed after 1990 may be the result of relocations of ethnic groups that have been repressed in their desired of returning to their fatherlands (e.g., Page 17.

(20) R. Soto. Growth and Employment in ECA. Germans in Poland). If average figures for the Eurozone or LAC are valid reference points (5% to 7%), then one can consider that out-migration has been “normal” in 13 ECA countries. However, double-digit out-migration levels require additional explanations. Political unrest –and outright civil violence—provides an additional explanation for some economies. While these explanations are valid, it seems that economic reasons also play a significant role (Kaczmarczyk and Okólski, 2005). In the early 1990s, the transition from communist rule to market economy led to massive restructuring and significant layoffs of public employees. Such transition revealed in several countries, particularly in Eastern Europe, a significant excess supply of labor with a concomitant increase in unemployment (a phenomenon largely absent prior to 1990). The rather low level of social protection led unemployed persons (often very young, particularly school leavers and relatively low-skilled) to seek jobs in western countries and in other countries of the region. A vigorous demand for such workers in Western Europe provided additional impetus to migration. Structural changes in trade patterns can also provide explanations for migration flows. Trade patterns were significantly affected during the transition; changes in relative prices and market size led trade within ECA to become less significant and trade with the EU more prominent. Such changes in trade structure led to changes in the structure of the demand for labor because they made certain abilities more demanded and others redundant. Workers with less demanded abilities had incentives to migrate.. 4.2.. Labor demand. The demand for labor in ECA, as elsewhere in the world, is largely pro-cyclical: in periods of expansion of economic activity employment also expands while in periods of recession it tends to contract. In the case of ECA economies the cyclical adjustment of the labor market has been superseded by the significant restructuring of the demand for labor as a result of the abandonment of socialist planning and, in particular the retrenchment of the public sector. In Table 6 I present annual growth rates of employment for the 1990s and 2000s. It can be seen that employment declined in the 1990s in almost all ECA countries in the EU+ group, being the only exceptions Romania, Slovenia, and Turkey. In the majority of cases these figures represent a systematic decline in employment while in only a few they are the product of a steep initial declines and posterior recovery in employment. The U-shaped trajectory is characteristic of Bulgaria, Hungary, Romania, Slovenia and the Slovak Republic. In the 2000s, there was some expansion in employment in all economies (except Romania) although it was significant only in the Slovak Republic and Turkey. On average, employment in this group of economies in 2011 was six percent lower than in 1990/91. The case of Slovenia is notable in that it has achieved significant growth in employment despite having a largely stationary population. The case of Romania is also noticeable but for the opposite reasons: despite coping better with the downturn in the early 1990s and achieving a recovery in employment in the early 2000s, this proved to be short-lived. By 2011, employment in Romania was 14% below its level of 2000. Page 18.

(21) R. Soto. Growth and Employment in ECA. The case of the Balkan economies is clear cut. In Albania and Bosnia and Herzegovina there is a steep initial decline in labor demand and a posterior recovery that is insufficient to return employment levels to pre-crisis levels. In the other economies, initial employment losses were not very significant and there was a continued expansion in demand afterwards. In the CIS, on the other hand, employment levels have evolved in markedly different ways. Employment grew in the 1990s only in countries where population was also growing (Azerbaijan, Kyrgyzstan, Tajikistan, Turkmenistan, and Uzbekistan). In fact, in these economies employment grew faster than population (compare Tables 3 and 4). Excepting Kazakhstan, in all other economies employment declined more or less at the same rate of the decline in population. In the 2000s, employment continued to grow above population growth in Azerbaijan, Kyrgyzstan, Tajikistan, Turkmenistan, and Uzbekistan. In the rest of the CIS economies, employment recovered steadily albeit very slowly at less than one percent per year. As a result, employment levels in 2011 were on average ten percent below the levels of 1990-91. When comparing the evolution of employment and GDP it is clear that there are significant lags in the long-term adjustment of the labor market; employment declined steadily in the 1990s even in economies where economic activity was recovering swiftly and did not necessarily expanded rapidly during the booming decade of the 2000s. Chronically high unemployment, as discussed below, in most economies reinforces this conclusion. Although part of the chronic unemployment could be the result of labor market mismatches (e.g., workers offering abilities no longer demanded by the private sector as a result of the transition to market economy), the persistence in unemployment indicates other deep-rooted problems including the inability to retrain the labor force (using active labor market policies), the aging of population (that makes retraining less cost-effective), and the lack of opportunities for rural workers. Contrary to countries in other regions, in ECA public employment is a significant component of the labor demand. Furthermore, it largely shaped the evolution of employment in the past two decades. Historically, employment in most ECA countries was mainly in state agencies and stateowned enterprises while the labor demand of the private sector was minimal; unemployment was also very low as the public sector operated as the employer of the last resort. In 1991-92, the share of public employment in ECA countries was on average slightly above 80%, as shown in Table 6. Only in Turkey the private sector commanded a significant share of the labor market with around 70%, but well below the levels of developed economies (85% to 90%) or emerging countries such as Chile (over 90%). The massive restructuring of the ECA economies –including privatization, divestiture, and downsizing—led to the emergence of a vigorous private sector labor demand and the retrenchment of the public sector to around 35% in 2010 on average. The retrenchment in public employment has been remarkably similar in the three country groups (EU+, CIS and Balkans), to the tune of twenty percentage points decade on decade, but the levels remain quite different. Countries in the CIS started with extremely high levels of public employment and, while it has declined, their economies remain dominated by public employment. Countries in the EU+ group started from lower levels of public employment and some are converging to Eurozone standards. Therefore, the observed evolution of employment reflects both market forces and administrative decisions on public-sector restructuring. However, there is no clear connection Page 19.

(22) R. Soto. Growth and Employment in ECA. between the size of the retrenchment in public employment and the response of the private sector, indicating that changes in total employment responded to more complex determinants than the mere substitution of one type of employer for another (as would have been by simply privatizing public firms). In some economies, public sector employment declined significantly and total employment expanded, as is the case in the Balkans (Albania and Macedonia), most of the CIS economies (except Belarus and Moldova) and the EU members (except Hungary, Lithuania and Romania).. Table 6 Total Employment, Public sector Employment, and Sectorial Employment Total Employment Annual Growth Rate (%). Public Sector Employment (as share of total employment, %). 1990-2000. 1990-2000. 2001-2011. 2001-2011. Sectorial Employment in 2008-2010 Agricul.. Industry. Services. Balkans Albania Bosnia Macedonia Montenegro Serbia. -0.6 -0.9a 0.6 0.7a 0.7a. 1.3 -0.6 1.4 -0.6 -1.5. 72.7 74.3 70.5 75.1 77.0. 47.5 47.9 44.5 57.0 55.6. 44.2 16.1 19.7 6.8 24.6. 18.9 25.0 31.3 20.9 25.7. 36.9 58.9 49.1 72.3 49.8. 16.8 12.8 40.0 9.1 18.9 20.6 19.7 28.4 9.3 14.0j 23.4 21.0k. 39.0 48.7 48.0 43.7 51.3 45.3 49.3 62.4 46.2 37.8j 60.7 45.0k. 35.0 29.1 39.0 32.4 31.4 27.3 26.0 31.1 30.1 38.4 33.7 26.1. 57.9 56.8 57.8 63.4 64.1 63.8 65.4 55.5 40.6 58.0 57.4 50.4. CIS Armenia -1.5 0.3 73.5 57.8 44.2 Azerbaijan 2.1 2.6 80.2 66.5 38.5 Belarus -1.0 -0.3 92.5 81.7 12.0 Georgia -0.3 0.2 82.4 60.3 47.1 Kazakhstan 0.9 2.1 74.6 44.1 29.8 Kyrgyzstan 1.5 1.7 80.6 58.1 34.0 Moldova -0.4 -2.5 74.4 51.4 31.1 Russia -0.5 0.7 71.5 48.0 9.2 Tajikistan 2.0 1.9 77.6 60.4 44.4 Turkmenistan 3.1 2.2 48.2j Ukraine -1.4 0.2 76.1 52.7 15.8 Uzbekistan 2.8 2.5 88.7 69.6 34.0k EU members, Croatia and Turkey Bulgaria -2.2 0.2 71.0 38.5 7.1 Croatia -1.5 0.3 68.0 48.5 14.1 Czech Rep. -0.7 0.4 46.9 32.6 3.1 Estonia -3.6 0.7 55.8 36.9 4.2 Hungary -1.0 -0.2 55.6 33.3 4.5 Lithuania -1.0 -0.1 69.4 44.6 8.8 Latvia -2.8 0.5 61.5 37.2 8.5 Poland -1.3 0.1 63.3 40.4 13.4 Romania 0.9 -1.3 63.1 39.1 29.3 Slovak Rep. -1.8 1.2 62.7 42.7 3.6 Slovenia 0.6 0.5 60.5 45.0 8.9 Turkey 1.3 1.6 29.1 22.5 23.4 Source: Own elaboration based on World Bank data, Life in Transitions and national statistics offices. Notes: (a) 1991-2000; (b) 1993-2000; (c) 1996-2000; (d) 1994-2000; (e) 2001-2010; (j) 2004; (k) 2006.. Finally, Table 6 contains information on the distribution of employment by sectors of economic activity. It can be seen the remarkable differences among countries: employment in the economies of the EU countries is primarily allocated to services –around 60 percent on average— Page 20.

(23) R. Soto. Growth and Employment in ECA. and to industry (30%) while those in the agricultural sector are usually less than 10%. On the contrary, in the CIS only Belarus, Russia and Ukraine resemble the EU economies, while in the rest countries agricultural employment is predominant (over 30% of total employment). In the Balkans, most countries are midway between EU and CIS members, except for Albania which is highly dependent on agriculture. By the same token, employment in the industrial sector is far more significant in the EU group (30%) than in the CIS (20%) or the Balkans (25%). The size of agricultural employment is of economic significance for understanding the evolution of labor markets in ECA. In most cases, agriculture and the rural world are viewed as lagging considerably in terms of income and employment opportunities and as the main source of migration of workers (the simple correlation between the size of the agricultural sector and the share of out-migrants in population in 2010 is a staggering 49%). Often stated reasons for this backwardness are the lack of agglomeration advantages, the low endowment of infrastructure and human capital, as well as the effects of structural changes in the economy towards a growing importance of services and globalization. In ECA these problems are compounded by the nature of reforms during the early stages of the transition to market economies: according to Petrick and Weingarten (2004) “in those countries which preserved the large-scale farming structures of the collectivist era, agriculture turned out to be more tailored to global competition than in countries where substantial restructuring into small (subsistence) farms took place” (page 9). Therefore, lack of dynamism in employment creation and a source for chronic unemployment and out-migration in the non-industrial ECA countries could be the results of inadequate agricultural reforms.. 4.3.. Labor market outcomes. As mentioned, data on market outcomes in ECA beyond employment and average labor productivity levels are quite difficult to obtain, particularly for the 1990s. With regards to nominal wages data collection is not systematic and suffers from methodological changes, it sometimes tends to reflect mostly public-sector wages (in some CIS countries) and price deflators needed to compute real wages are unavailable for economies that suffered from high or hyper-inflation. Furthermore, data on wages tend to represent public sector wages rather than a weighted average of the latter and wages in the private sector. Data on hours are absent for most CIS and Balkan economies and when available it usually refers only to the period 2000-2010. Table 7 collects the available information. Data on average labor productivity correspond to total GDP divided by the number of employed workers. It can be seen that productivity levels declined quite significantly in the CIS in the 1990s: the moderate decline in Belarus, Kazakhstan and Armenia contrasts with the massive collapse in Azerbaijan, Georgia, Kyrgyzstan, Moldova, Tajikistan, Turkmenistan, and Ukraine. No other country in ECA experienced as such collapse with the only exceptions of Lithuania and Serbia. The EU members suffered from less significant drops in the early 1990s and productivity recovered in the second part of the decade to return to the 1990 levels. The recovery period is characterized by productivity gains of around 3% to 5% per year in countries that are members of the EU and the Page 21.

(24) R. Soto. Growth and Employment in ECA. Balkans. This is a relatively high growth rate for international standards. In the CIS, however, average productivity growth is much higher on average and quite heterogeneous among countries reflecting both the lasting effects of the lost decade of 1990 and their ability to implement progrowth policies.. Table 7 Average Labor Productivity, Real Wages, and Hours Worked Average Labor Productivity Annual Growth (%) 1991-1995. 2001-2011. -2.4 -7.3 -4.9 -15.8. 3.6 4.4 1.1 4.1 5.1. Real Wages Annual Growth Rate (%) 1991-2000. 2001-2011. Average Hours Worked per Week 2000. 2010. 34.8 -. 42.2 41.4 42.4 -. Armenia -13.3 7.3 10.6b 10.5 43.5 Azerbaijan -21.3 10.7 1.0b 11.8 23.6 39.3 Belarus -8.7 7.6 Georgia -23.9 6.1 15.0 15.6 Kazakhstan -7.8 6.0 -0.6b 8.9 35.0 11.1 34.8 Kyrgyzstan -14.3 2.4 Moldova -19.0 7.9 11.3 28.1 11.4 38.1 Russia -9.4 4.0 Tajikistan -25.4 6.1 20.7 Turkmenistan -17.8 11.2 12.8 32.5 Ukraine -16.3 4.2 -2.5b Uzbekistan -7.6 4.4 EU members, Croatia and Turkey Bulgaria -3.5 3.6 -4.6 4.2 40.6 Croatia -5.8 2.1 22.5b 1.7 40.6 Czech Rep. -0.1 2.6 4.1b 2.9 43.0 Estonia -4.4 3.6 6.1b 4.5 40.5 2.0 40.9 Hungary -1.2 2.1 -0.7 Lithuania -10.8 4.6 -4.4 3.0 39.3 Latvia -8.9 3.3 12.3 5.3 42.4 Poland 0.0 3.9 4.6 1.2 40.6 Romania -4.0 5.5 6.6 9.5 39.4 Slovak Rep. -1.4 3.5 4.2d 2.3 41.6 Slovenia -1.4 1.9 -1.4 0.8 41.0 Turkey 0.0 2.6 -1.8 -2.8e 49.8 Source: Own elaboration based on World Bank data, Life in Transitions and national statistics offices. Notes: (a) 1991-2000; (b) 1993-2000; (c) 1996-2000; (d) 1994-2000; (e) 2001-2010; (j) 2004; (k) 2006.. 38.5 37.9 39.5 36.0 31.4 38.0 35.3 -. Balkans Albania Bosnia Macedonia Montenegro Serbia. 2.7 11.3c 3.0 -. 6.5 5.7 2.0 8.5. CIS. 40.5 39.0 40.4 38.4 39.6 38.4 38.4 39.6 39.2 39.5 38.6 47.4. The changes in real wages during and after the transition do not conform to classical market responses to changes in economic activity. The massive change in relative prices brought upon by market liberalization, privatization and, in particular, the drop in GDP and the massive public sector retrenchment would have called for a significant decline in real wages. Although the data are scarce Page 22.

(25) R. Soto. Growth and Employment in ECA. for the early 1990s, the evidence indicates that real wages dropped significantly in some countries with the unveiling of the economic crisis. However, real wages actually increased in economies that were experiencing severe contractions in both output and employment, such as Albania, the Czech Republic, Romania or Estonia. A similar behavior was observed in the late 1990s in other economies when real wages grew quite fast even though employment was declining at the same time: in Armenia, Belarus, Georgia, and Poland real wages grew at double-digit levels in the period 1995-2000 while employment was declining at significant rates. It has been argued that nominal wages are computed largely on the basis of public sector wages and that the latter had been adjusted upwards during the crisis for political reasons and with disregard of labor market conditions. Still, as noted, the public sector was –and in some countries continues to be—the main employer so to that extent wage indices are representative of the market conditions, distorted as they may be. Growth in real wages in the recovery period seems to be more closely related to sustained growth in productivity levels in the majority of countries. As shown in Figure 3, there is a clear positive relationship between the average growth in labor productivity in the 2000s and the increases in real wages. Two conclusions can be highlighted from these data. First, real wage growth seems to be more contained and in line with productivity increases in the EU+ economies than in the CIS, with the only exception of Romania. In the latter real wages have increased at around 10% per year in the period 2000-2011 yet economic growth has been only around 5% per year. In the CIS, there is a group of economies where real wages have grown quite fast in the 2000s but where average productivity levels have also expanded in a vigorous manner. This group includes oil and gas producers and exporters (Azerbaijan, Russia, and Kazakhstan) as well as two Balkan economies, Serbia and Bosnia and Herzegovina. The other group exhibits extremely high growth rates in real wages that are in excess of labor productivity increases and general growth: in Belarus, Georgia, Moldova and Tajikistan real wages increases exceed 15% per year yet productivity gains do not exceed 5%. Perhaps one of the most intriguing outcomes of labor markets in ECA –in particular when confronted with the evolution of real wages—is the persistence of unemployment in some economies as shown in Table 8. Unemployment statistics are controversial in ECA countries as in some cases they refer to administrative data (e.g., those registered in unemployment agencies) while in others the data come from labor force surveys. Notwithstanding the limitations of the data, it can be seen that unemployment levels remained stubbornly stagnant in the period 2000-2008 in several countries despite de fact that these economies were growing at very high rates; these include all the Balkan states and most CIS economies. Only in the EU group unemployment rates declined from relatively high levels in 2000 and converged by 2008 to what can be termed as frictional unemployment. In the CIS, four economies (Kazakhstan, Moldova, Russia, and Turkmenistan) reduced unemployment significantly and leveled with the EU group. Four other CIS economies continued to display double-digit unemployment rates, despite the very fast pace of economic growth (over 6% per year in the period 2000-2011).. Page 23.

(26) R. Soto. Growth and Employment in ECA. Figure 3. Own elaboration based on data from the World Bank.. It is, to some extent, difficult to reconcile high, chronic unemployment levels as those observed in some CIS economies or the Balkans with evidence of very high real wage growth of the period 2000-2011 (e.g., Armenia, Georgia, Tajikistan). Naturally, it may reflect a statistical artifact in that wage data may cover only a small fraction of the labor force –typically, public sector workers—and this group may not represent adequately the market stance. However, such explanation requires a rationale to justify why governments would like to raise wages far above national productivity gains. Most likely the chronically high rate of unemployment reflects the existence of lowemployability workers that tend to remain unemployed for long periods of time while real wage indices represent the evolution of employment opportunities for those who are highly employable. Long unemployment duration impairs the human capital of the unemployed and discourages job search which, in turn, reinforces the stigmatization of those unemployed for a long time as low quality workers. In this framework, labor rigidities are bad not only because they restrain labor demand from increasing but also because low and delayed hiring rates imply an increase in the duration of unemployment.. Page 24.

(27) R. Soto. Growth and Employment in ECA. Table 8 Unemployment Rates, selected years (%) 2000. 2004. 2008. Balkans Albania 14.6 14.3 13.0 Bosnia 25.2 29.1 23.9 Macedonia 32.3 37.2 33.7 Montenegro 15.6 22.5 16.8 Serbia 12.6 18.5 13.6 CIS Armenia 33.7 33.9 28.6 Azerbaijan 7.2 9.0 6.0 Belarus 10.1 9.8 9.7 Georgia 10.8 12.6 16.5 Kazakhstan 12.8 8.4 6.6 Kyrgyzstan 7.5 8.6 8.2 Moldova 8.5 8.2 4.0 Russia 10.6 7.7 6.3 Tajikistan 11.2 11.3 11.5 Turkmenistan Ukraine 11.6 8.6 6.4 Uzbekistan EU members, Croatia and Turkey Bulgaria 16.2 12.0 5.6 Czech Rep. 8.8 8.3 4.4 Estonia 13.1 9.9 5.5 Croatia 16.1 13.7 8.3 Hungary 6.4 6.1 7.8 Lithuania 14.2 9.9 7.4 Latvia 16.0 11.3 5.9 Poland 16.1 19.0 7.1 Romania 7.0 7.7 5.8 Slovak Rep. 18.8 18.1 9.6 Slovenia 7.2 6.2 4.4 Turkey 6.5 10.8 11.0 Source: The ILO's KILM database 7th edition (2012) and World Bank (2012).. 4.4.. 2009. 2010. 13.8 24.1 32.2 19.1 16.1. 13.7 27.2 32.0 19.7 19.2. 30.5 6.0 10.1 16.8 6.6 8.4 6.4 8.4 11.7 8.8 -. 29.3 6.3 9.7 16.4 5.8 8.8 7.5 7.5 11.6 8.5 -. 6.8 6.7 13.7 9.0 10.0 17.2 13.7 8.2 6.9 12.0 5.8 14.0. 10.2 7.3 16.9 11.8 11.2 18.7 17.8 9.6 7.3 14.4 7.3 11.9. Adjustment in the extensive/intensive margins. Adjustments in the labor market occur in two dimensions or margins. First, the number of workers can be adjusted as needed for production purposes. This is usually called the extensive margin. Second, firms can adjust the intensity of the effort (number of hours worked) by each worker. This is usually called the intensive margin. Modern market economies usually employ both margins as the optimal economic response to changes in demand and the business cycle. The reason for using both margins is that the hiring and firing workers in response to production needs tends to be a costly process. In addition to the entry and exit costs (recruiting costs, severance payments and the like), there is uncertainty about the true productivity of new workers and about the future availability of equally qualified personnel once a downturn is over. Page 25.

(28) R. Soto. Growth and Employment in ECA. In principle, there is a wealth of information arising from the study of both margins. For example, when hiring/firing costs are too high or when firm-specific human capital is an important component of costs, then adjustments would tend to occur in hours more than in workers as employers would find it in their advantage to avoid laying-off workers. On the contrary, in economies where job schedules are rigid or part-time employment is penalized and where human capital is not important, then adjustments would tend to rely more on the number of workers. Adjustments in the extensive margin are also limited by the availability of manpower in some ECA countries. As mentioned above, population growth is in general extremely low and in some countries it is negative, while participation rates tend to be relatively high and stable. In principle, this limits the capacities of an economy to use its extensive margin during an economic boom (increase employment) and forces the adjust more on the intensive margin (increase hours per worker). Likewise, the scarcity of workers gives incentive to employers to retain manpower during downswings and rely on adjusting hours. In ECA, economies can also recourse to migration to adjust for lack of national workers –as has been the case in the past—but this is of course slow and quite costly. I therefore expect the use of the extensive margin to be more pronounced in the seven economies identified above, while in the rest adjustments in the intensity should be the norm. The availability of data on hours worked for ECA countries –which is needed to compute the changes in the intensive margin—is unfortunately very limited and to some extent unreliable (for example, underemployment is usually not measured). The underlying surveys used to construct these series, whether sampling establishments or the labor force, are not uniform across countries and, in some cases, for the same country at two different dates. In general, there are no data available before 2000, thus limiting our comparative analysis to the last decade, and only for a handful of countries. Information on the Balkan economies and the CIS is much weaker and scarcer. In Table 9 I present the available data on the growth of total hours worked and its decomposition into the growth in employment (extensive margin) and the growth in the average number of hours worked by workers (the intensive margins). It can be seen that the information on total hours is available for the EU economies but quite scarce for the countries in the CIS and the Balkans. Six economies do not provide any information on hours worked (Georgia, Azerbaijan, Tajikistan, Turkmenistan and Uzbekistan). In the latter four of these economies, relatively high population growth would indicate that there is space for adjusting the extensive margin. Prior to the recent global economic crisis –in the period 2000-2004—the changes in total hours worked largely reflect the evolution along the extensive margin in eight economies: Albania, Belarus, Bulgaria, Estonia, Hungary, Poland, Russia, and Turkey. I consider any change in the number of hours worked by a worker below 0.5% to be negligible (in a workweek of 35 hours it would amount to less than two hours). In two economies –Kyrgyzstan and Ukraine—there was an increase in both employment and hours worked per worker. In Moldova and Romania, somewhat paradoxically, hours per worker increased at the same time that employment was declining further fueling unemployment issues. Finally, in several EU+ countries worked hours per worker declined by around 1% per year but while in Lithuania, Latvia, Slovenia and the Slovak Republic the. Page 26.

(29) R. Soto. Growth and Employment in ECA. reduction in effort per worker was partially compensated by higher employment levels, in the Czech Republic it was purely an intensive margin adjustment.. Table 9 Adjustments in the Extensive and Intensive Margins of the Labor market Annual growth in total hours worked (%) 2000-04. 2004-08. 2008-10. Annual growth in employment (%) 2000-04. 2004-08. 2008-10. Annual growth in hours worked per worker (%) 2000-04. Balkans Albania 1.2 1.1 1.7 0.5 Bosnia -2.5 0.9 -1.6 -1.5 Macedonia 1.2 -1.8 3.9 2.6 Montenegro -2.5 -2.0 1.7 -2.9 Serbia -2.0 1.7 -7.8 CIS Armenia -2.6 -0.5 1.0 0.9 Azerbaijan 2.1 3.5 2.6 Belarus 0.2 -1.1 -0.2 -0.8 -0.6 Georgia -0.1 -0.2 1.4 2.1 2.3 2.7 Kazakhstan 2.3 Kyrgyzstan 2.5 1.4 2.6 1.4 Moldova -0.1 -1.7 -2.4 -2.1 -5.1 1.4 -0.9 0.6 1.4 -0.4 Russia 0.8 Tajikistan 1.2 2.0 2.5 Turkmenistan 2.1 2.2 2.1 1.2 0.4 0.9 -1.1 Ukraine 2.1 Uzbekistan 2.6 2.9 3.0 EU members, Croatia and Turkey Bulgaria -0.1 3.8 -5.1 -0.1 3.5 -4.6 Czech Rep. -1.1 1.3 -1.3 -0.1 1.6 -1.2 Estonia 1.5 2.2 -8.8 1.4 2.6 -6.7 Croatia 1.0 -3.1 1.1 1.4 -2.5 Hungary -0.1 -0.2 -1.9 0.4 -0.2 -1.3 Lithuania -0.3 2.0 -7.0 0.6 1.1 -6.0 Latvia 1.4 1.5 -10.0 2.2 2.6 -8.8 Poland -1.0 3.7 -0.5 -1.0 3.9 0.2 Romania -2.8 -0.4 -1.3 -3.6 0.2 -0.8 Slovak Rep. -0.1 3.7 -2.1 0.9 2.9 -2.5 Slovenia 0.6 1.3 -0.9 1.5 1.0 -1.1 Turkey -0.8 1.9 2.6 -0.5 2.2 3.3 Source: Own elaboration based on data from World Bank (2012), ILO (2012) and Eurostat (2012).. 2004-08. 2008-10. -0.1 -1.0 -1.4 0.3. -3.6 0.5. -0.3 0.0. 1.1 2.3 0.2. 0.0. 1.7. 0.4. 0.0 -1.0 0.1. 0.3 -0.4 -0.3 -0.4 -0.1 0.9 -1.1 -0.2 -0.7 0.7 0.3 -0.2. -0.5 -0.9 -0.8 -0.1 0.8 -1.0 -0.9 -0.2. -0.5. -0.6 -0.1 -2.2 -0.6 -0.6 -1.0 -1.2 -0.8 -0.5 0.4 0.2 -0.6. Adjustments in the labor market in the booming years of 2004 to 2008 were comparable to those in the previous period. Twelve economies adjusted primarily along the extensive margin either expanding or contracting employment but refraining from changing working hours (Belarus, Kazakhstan, Russia, Ukraine, Bulgaria, the Czech Republic, Estonia, Croatia, Hungary, Poland, Slovenia, and Turkey). In Lithuania both margins contributed equally to the increase in total hours worked demanded by firms while in the Slovak Republic there was more reliance in the extensive margin. In Armenia, Romania and Latvia worker’s efforts declined while employment increased.. Page 27.